spam.chol.NgPeyton-class.Rd
Result of a Cholesky decomposition with the NgPeyton
method
Objects are created by calls of the form chol(x,method="NgPeyton", ...)
and should not be created directly with a
new("spam.chol.NgPeyton", ...)
call.
At present, no proper print method is defined. However, the factor can
be transformed into a spam
object.
signature(x = "spam.chol.NgPeyton")
: Transform the factor
into a regular matrix.
signature(x = "spam.chol.NgPeyton")
: Transform the factor
into a spam
object.
signature(r = "spam.chol.NgPeyton")
: solving
a triangular system, see solve
.
signature(l = "spam.chol.NgPeyton")
: solving
a triangular system, see solve
.
signature(x = "spam.chol.NgPeyton")
: Coerce the factor into a vector.
signature(x = "spam.chol.NgPeyton")
:
Calculates the determinant from the factor, see also det
.
signature(x = "spam.chol.NgPeyton")
: Extracts the
diagonal entries.
signature(x = "spam.chol.NgPeyton")
: Retrieve the
dimension. Note that "dim<-"
is not implemented.
signature(x = "spam.chol.NgPeyton")
: Transformation
to a spam
object and display, see also display
.
signature(x = "spam.chol.NgPeyton")
: Transformation
to a spam
object and display, see also image
.
signature(x = "spam.chol.NgPeyton")
: Retrieve the
dimension. Note that "length<-"
is not implemented.
signature(x = "spam.chol.NgPeyton")
:
Retrieves the ordering, in ordering
.
signature(x = "spam.chol.NgPeyton")
: Short description.
signature(object = "spam.chol.NgPeyton")
: Short description.
signature(object = "spam.chol.NgPeyton")
:
Description of the factor, returns (as a list) nnzR
, nnzcolindices
,
the density of the factor density
, and fill-in ratio
fillin
. For the use of the first two, see ‘Examples’
in chol
.
signature(x = "spam.chol.NgPeyton")
: Transformation
to a spam
object and transposition.
signature(x = "spam.chol.NgPeyton")
: Returns
x
unchanged.
It is not possible to directly change the length, dimension and the
diagonal entries of a "spam.chol.NgPeyton"
object.
Ng, E. G. and B. W. Peyton (1993), "Block sparse Cholesky algorithms on advanced uniprocessor computers", SIAM J. Sci. Comput., 14, pp. 1034-1056.
print.spam
ordering
and chol
x <- spam( c(4,3,0,3,5,1,0,1,4),3)
cf <- chol( x)
cf
#> (Upper) Cholesky factor of dimension 3x3 with 5 (row-wise) nonzero elements (
#> 2.42 Kb).
#> (The object is supposed to be used with: 'as.spam', 'backsolve', 'forwardsolve', etc.)
#>
#> Class 'spam.chol.NgPeyton'
as.spam( cf)
#> [,1] [,2] [,3]
#> [1,] 2 0 0.500000
#> [2,] 0 2 1.500000
#> [3,] 0 0 1.581139
#> Class 'spam' (32-bit)
# Modify at own risk...
slotNames(cf)
#> [1] "entries" "colindices" "colpointers" "rowpointers" "dimension"
#> [6] "pivot" "invpivot" "supernodes" "snmember" "memory"
#> [11] "nnzA"